The Measure of Balance in Sitting in Stroke Rehabilitation Prognosis Karl J. Sandin, MD, and Barry S. Smith, MD To test the hypothesis that recovering stroke patients with initially good sitting balance or those who develop good sitting balance during rehabilitation have better Barthel Index-based functional assessment outcomes than recovering stroke patients with poor sitting balance, we prospectively studied 24 of 25 consecutive stroke patients admitted to a tertiary-care hospital rehabilitation unit Sitting balance, a prerequisite for most functional activities, was scored using a four-point scale evaluating static and dynamic sitting balance on admission to the rehabilitation unit and weekly until discharge. Functional status was assessed using the Barthel Index 4 or 12 weeks after the stroke. We found a strong positive correlation between Barthel Index score and each weekly sitting balance score. Multiple evaluations over time identified those patients whose sitting balance improved during rehabilitation in our unit; after grouping the patients into those with normal, improved, and poor sitting balance, we found a significant difference in the Barthel Index scores among the three groups. The group of patients whose sitting balance improved had higher Barthel Index scores than the group whose sitting balance did not improve. Ongoing evaluation of sitting balance can be helpful in anticipating functional status at discharge in stroke patients undergoing comprehensive rehabilitation. (Stroke 1990^1:82-86)
hysiatrists have struggled to define the relation between various aspects of functional status and overall rehabilitation outcome in stroke patients. Many studies have correlated patient characteristics present at the initial examination with long-term functional improvement; other studies have been unable to document a sufficiently high correlation to precisely predict outcome. Anderson et al1 showed that perceptual loss, low motivation, confused and disoriented thinking, withdrawn and apathetic behavior, an extended time since stroke onset, previous stroke, low blood pressure, and an extended period of unconsciousness at the time of the stroke are important variables related to improvement. Lehmann et al2 showed that family income and family involvement supporting the patient predicted discharge disposition. There is a paucity of data concerning the value of specific, ongoing functional assessment techniques in identifying those stroke victims who will do well in inpatient rehabilitation units. We report the results of a prospective study of 24 of 25 consecutive stroke patients admitted to the rehabilitation unit of an acute-care hospital. We From the Department of Physical Medicine, Baylor College of Medicine (K.J.S., B.S.S.) and the Department of Physical Medicine and Rehabilitation, St. Luke's Episcopal Hospital (B.S.S.), Houston, Texas. Address for correspondence: Dr. Sandin, Department of Physical Medicine, Baylor College of Medicine, 1333 Moursund, A-221, Houston, TX 77030. Received April 10, 1989; accepted August 1, 1989.
serially evaluated sitting balance to test the hypothesis that recovering stroke patients with initially good sitting balance or those who develop good sitting balance during rehabilitation do better on a Barthel index-based functional assessment of outcome than recovering stroke patients with consistently poor sitting balance. Subjects and Methods
Twenty-five patients who had had a stroke were admitted to the rehabilitation unit from May 2 to September 5, 1988. Twenty-four were referred from St. Luke's Episcopal Hospital, a large tertiary-care teaching hospital emphasizing medical and surgical treatment of cardiovascular disease and the site of the rehabilitation unit. The other patient was referred from the outpatient clinic of our department. One patient had a hip fracture in addition to a stroke and he was required to wear a thigh abduction splint at all times. Because his sitting balance could have been affected by his fracture, his stroke, or his postoperative orthopedic equipment, this patient was deleted from the cohort. The same physiatrist obtained informed consent from each patient or a family member, tested the patient's sitting balance, and assessed the patient's Barthel Index score (see below). Sitting balance was tested using the standard technique for evaluating static and dynamic sitting balance.3 The patient sat on the side of a hospital
Downloaded from http://stroke.ahajournals.org/ by guest on February 2, 2015
Sandin and Smith Serial Evaluation of Sitting Balance
bed, feet on the floor, back unsupported, and hands on the lap. If the patient could hold this position without assistance for 15 seconds, he was nudged by the physiatrist anteriorly, posteriorly, and laterally using approximately 5-10 foot-pounds of force. The physiatrist guarded the patient from falling with his free hand. The patient's sitting balance was scored as 4, normal: able to perform the above testing without any physical assistance; 3, good: able to maintain a static position without difficulty but requiring assistance in righting from the hemiplegic side; 2, fair: able to maintain a static position without difficulty but requiring assistance in all righting tasks; or 1, poor: unable to maintain a static position. Sitting balance was evaluated on admission to the rehabilitation unit and every week while the patient was in the hospital. The Barthel Index,4 an ordinal scale that comprises scores for feeding, mobility, personal care, ambulation or wheelchair skills, bowel and bladder abilities, and dressing skills, was used to assess each patient's rehabilitation outcome. The Barthel Index was selected because of its ease of administration, its proven reliability, and its good track record in the functional evaluation of patients with stroke.5"7 The Barthel Index score was generated by physiatric examination in conjunction with rehabilitation nursing evaluation exactly 4 weeks after the onset of hemiparesis due to stroke in 21 of the 24 patients and 12 weeks after stroke onset in three patients with extremely long hospital stays before admission to the rehabilitation unit. Each patient's discharge disposition (to home without or with assistance, to retirement home, to extended-care facility/nursing home, or back to the acute-care hospital) was recorded. The percentage of the cohort with given demographic and clinical characteristics was calculated. The mean, standard deviation (SD), range, median, and mode were calculated for the time variables age, time from stroke onset to admission to rehabilitation unit, time to initial sitting balance evaluation, and time in rehabilitation unit. The correlation coefficient (r) for sitting balance score versus Barthel Index score was calculated and evaluated using the two-tailed t test. The mean±SD Barthel Index scores for defined groups of patients were calculated and compared using the two-tailed t test. Results Of the 24 patients, 16 (66.7%, five men and 11 women) had left hemiparesis and eight (33.3%, four men and four women) had right hemiparesis; 22 (91.7%) strokes were thromboembolic and two (8.3%) consisted of capsular hemorrhage that resulted in left hemiparesis. This was the first stroke in 23 patients. The patient who had already had several minor strokes was functioning normally at the time of the stroke under consideration. The mean±SD age of the 24 patients was 70.7±8.3 (range 51-82) years. The demographic and clinical data are summarized in Table 1.
The 24 patients were admitted to the rehabilitation unit an average of 17.0 (SD 12.9, range 3-60) days after their stroke (Table 1); the median and mode were 14 days. Sitting balance was initially evaluated an average of 3.5 (SD 2.9, range 0-11) days after admission to the rehabilitation unit; the median was 3 days. The time from the onset of stroke to the initial sitting balance evaluation averaged 20.5 (SD 12.8) days; the median was 16.5 and the mode 16 days. The 24 patients underwent a median of two (range 1-4) sitting balance evaluations while in the rehabilitation unit (Table 1). Table 2 gives the frequency distributions of the sitting balance scores for each evaluation. Because the Barthel Index scores for three patients were obtained 12 weeks after the stroke, the statistical analyses were performed first by including and then by omitting the results from those three patients. Because there was no significant difference in the results between analyses, we report the analysis of all 24 patients. The Barthel Index score averaged 75 (SD 17, range 35-95); the median and mode were 80. Of the 24 patients, 18 (75%) were discharged to home with help, including stand-by assistance, and one (4.2%) was discharged to independent home care. Two patients (8.3%) were discharged to extended-care facilities and one (4.2%) to a retirement home. Two patients (8.3%) returned to the acute-care hospital, where one later died. Correlation analysis of the Barthel Index score and the first sitting balance scores yielded r=0.699; (p